Table 2.

Estimates of log(OR) in Simulation II

P*	Method	β1 = 0.8					β1 = 0
		Bias†	ESE‡	ASE§	RMSE¶	CP(%)**	Err(%)††
0.1	Full	0.013	0.088	0.082	0.089	0.94	5.6
	Hlf DL	−0.061	0.082	0.079	0.102	0.85	5.3
	Deletion	0.014	0.098	0.093	0.099	0.94	5.6
	MI	0.003	0.111	0.114	0.111	0.95	4.6
	ROS	−0.007	0.086	0.083	0.086	0.95	5.2
	Bayesian	0.004	0.088	0.083	0.088	0.94	5.9
0.3	Full	0.009	0.086	0.082	0.086	0.94	6.5
	Half DL	−0.193	0.067	0.068	0.205	0.20	6.0
	Deletion	0.017	0.124	0.115	0.125	0.93	5.1
	MI	−0.045	0.166	0.170	0.172	0.93	3.2
	ROS	−0.086	0.080	0.087	0.117	0.83	3.6
	Bayesian	−0.020	0.089	0.085	0.091	0.93	7.0
0.5	Full	0.010	0.083	0.082	0.084	0.95	5.5
	Half DL	−0.257	0.060	0.061	0.264	0.03	6.0
	Deletion	0.028	0.154	0.150	0.156	0.95	5.7
	MI	−0.141	0.242	0.241	0.280	0.88	2.4
	ROS	−0.214	0.072	0.091	0.225	0.32	2.2
	Bayesian	−0.039	0.090	0.090	0.098	0.91	5.8

^*P is the proportion of DL

^†Bias is defined as β-β

^‡ESE is the empirical standard error

^§ASE is the average standard error

^¶RMSE is the root mean square error

^**CP is the coverage probability for 95% HPD

^††Err is the type I error rate under null hypothesis.